Paediatrica Indonesiana

VOLUME 44 March - April • 2004 NUMBER 3-4

Original Article

Correlation between leukocyte aggregation score of

cerebrospinal fluid and bacterial meningitis in children
Meitha PE Togas, MD; Nurhayati Masloman, MD

ABSTRACT could early distinguish the etiology of meningitis in

Background Bacterial meningitis is one of life-threatening dis-
eases and carries a risk of sequelae in affected children. In terms
of cost and rapid differentiation between bacterial and non-bacte-
rial meningitis, several tests have been proposed.
Objective This study aimed to determine the use of leukocyte
aggregation score (LAS) of cerebrospinal fluid (CSF) in diagnos-
ing bacterial meningitis.
Methods A prospective analytic study was done from October 2001
to July 2002 in the Department of Child Health, Medical School,
Sam Ratulangi University/ Manado General Hospital. Children pre-
senting with symptoms of meningitis, aged between 28 days and
13 years were enrolled. LAS was counted in percentage. Regres-
sion analysis was used to determine the correlation between LAS
and diagnosis of bacterial meningitis.
Results CSF examinations were done on 35 meningitis patients.
Three patients were excluded. The remaining 32 patients com-
prised of 11 with bacterial meningitis and the other 21 with non-
bacterial meningitis. The mean of LAS in bacterial meningitis was
significantly higher than that of non-bacterial meningitis (p<0.001).
The cut off value of LAS to diagnose bacterial meningitis was
12.35%.
Conclusion LAS may be used as a fast and simple alternative
diagnostic tool to confirm the diagnosis of bacterial meningitis
[Paediatr Indones 2004;44:61-65].
Keywords: bacterial meningitis, leukocyte aggrega-
tion score, cerebrospinal fluid
infant and children.1-3 Yet, there are many reports
about the efforts to diagnose bacterial meningitis as
early as possible, such as CSF lactate level,4 C-reactive
protein (CRP), 5-7 lactoferrin, α-1 antitrypsin,
immunoglobulin GA,7 tumor necrosis factor (TNF), 8-
10 interleukin (IL) 1α,10-11 IL-6, 12 granulocyte colony
stimulating factor,13 bacterial antigen detection test,14
and CSF leukocyte aggregation score (LAS).15-16 CSF
LAS examination is simple, fast, not expensive, and
does not need sophisticated equipment or skills.16
The objective of this study was to evaluate the
use of LAS examination in establishing early diagno-
sis of bacterial meningitis and to compare it with stan-
dard examinations.
Methods
A prospective analytic study was conducted at the
Department of Child Health, Manado General
Hospital from October 2001 to July 2002. The
inclusion criteria were infants and children aged
between 28 days and 13 years with symptoms and signs

B
acterial meningitis is one of life-threatening
neurologic diseases, especially in neonatal
and childhood, that carries a high risk of
From the Department of Child Health, Medical School,
neurologic sequelae and still has high Sam Ratulangi University, Manado General Hospital, Manado.
mortality.1-3 Early diagnosis and prompt treatment are
important in the management of bacterial meningitis. Reprint requests to: Nurhayati Masloman, MD, Department of Child
Health, Medical School, Sam Ratulangi University, Manado General
But early symptoms and signs are not specific, mainly Hospital, Manado, Indonesia. Tel. 62-431-821652, Fax. 62-431-859091,
in infant. There had been no single examination that Email: ikarsup@wasantara.net

Paediatrica Indonesiana, Vol. 44 No. 3-4 • March - April 2004 • 61

 Paediatrica Indonesiana

of meningitis who were admitted to our department.
The exclusion criteria were infants or children who
had contraindication of lumbar puncture (LP) (shock,
severe general condition, skin infection around the
location of LP, space occupying lesion, hydrocephalus,
blood dyscrasia, hyperventilation, irregular respiration,
and apnea), blood-contaminated CSF whether grossly
or microscopically, and other conditions such as
meningismus, malignancy, collagen-vascular
syndromes, toxin exposure, and focal infection of the
central nervous system (i.e., brain abscess,
parameningeal infection, and subdural empyema).
Informed consent was obtained from the parents.
The diagnosis of bacterial meningitis is based on
symptoms and signs of meningitis i.e., fever, seizure,
and unconsciousness, plus at least one of the follow-
ings: 1) biochemistry and cytology results of CSF ex-
amination were opalescent to turbid in color, Nonne
(+), Pandy (+), protein >40 mg/dL, glucose <40
mg/dL, leukocyte count 1,000-10,000/mm3 with pre-
dominant PMN, and positive Gram staining; 2) posi-
tive bacterial growth on culture.
After LP had been performed, the CSF obtained
was divided into 2 tubes. One tube containing 1-2
mL CSF was sent to Prodia Clinical Laboratory for
biochemistry, cytology, and Gram staining examina-
tion within 90 minutes. The second one containing 1
Fisher’s exact test was used to compare two categori-
cal variables.
Results
From October 2001 to July 2002, CSF examinations
were conducted on 35 patients with symptoms and
signs of meningitis who were admitted to the
Department of Child Health, Medical School, Sam
Ratulangi University, Manado General Hospital.
Three patients were excluded due to technical faults
i.e., one without Pandy and protein level examination
results, one with biochemistry, cytology, and Gram
staining examination performed in more than 90
minutes after LP, and the other one with insufficient
CSF sample for bacterial culture. The remaining 32
patients comprised of 11 patients with bacterial
meningitis and 21 with non-bacterial meningitis.
The characteristics of the patients showed that
most were in the age of <2 year-old. The ratio of male
to female was 1:1.
Table 1 shows that positive bacterial culture was
found mostly in the age of <2 years (6 from 11 pa-
tients).
TABLE 1. DISTRIBUTION OF PATIENTS WITH POSITIVE CULTURE

mL CSF was sent as soon as possible to the Depart- RESULTS ACCORDING TO BACTERIAL ETIOLOGY AND AGE GROUP

ment of Microbiology, Medical School, Sam Ratulangi Bacterial etiology Age Group
University for bacterial culture. The procedure to < 2 yrs 2-5 yrs > 5 yrs Total
make LAS slide was as follows: one to two drops of Enterobacter aerogenes 1 2 0 3
CSF was put on an object glass, then it was laid on a Staphylococcus epidermidis 3 0 0 3
Alkaligenes faecalis 0 1 1 2
45o declined base for 3-4 seconds. The CSF was pre- Citrobacter difersus 1 0 0 1
cipitated by gravitation, leaving a thin layer. Then, it Acinetobacter sp. 0 1 0 1
was dried on a horizontal position, fixed by absolute Serratia marsecen 1 0 0 1
Total 6 4 1 11
methanol and stained with hematoxyllin-eosin. Two
skillful pathologists from the Department of Clinical
Pathology, Medical School, Sam Ratulangi University,
read it under a light microscopes without knowing Table 2 shows the mean and standard deviation
either the patient’s condition or the results of CSF of age, protein, glucose, leukocyte, and LAS of CSF
biochemistry, cytology, and culture. The score was according to the diagnosis. It was obvious that only
then given in percentage. the mean of LAS that was significantly higher in bac-
Data were analyzed descriptively and analytically. terial meningitis than that in non-bacterial meningi-
Regression model was used to determine the correla- tis (p<0.001).
tion between two variables, with formula: y = Bx+E The correlation between diagnosis and the results
(y for dependent variable i.e., diagnosis of bacterial of CSF examinations including biochemistry, cytology,
meningitis; x for independent variable, i.e., LAS; B and culture is shown in Table 3; only LAS that was
for regression coefficient; and E for other variable). significantly correlated to bacterial meningitis. Accord-

62 • Paediatrica Indonesiana, Vol. 44 No. 3-4 • March - April 2004

 Meitha PE Togas et al: Correlation between leukocyte aggregation score of CSF and bacterial meningitis

TABLE 2. MEAN AND STANDARD DEVIATION (SD) OF VARIABLES

ACCORDING TO DIAGNOSIS.

Diagnosis
Variables Bacterial Non-bacterial p
Meningitis Meningitis
(mean+SD) (mean+SD)
Age (years) 2.77+2.32 4.06+3.62 0.296
Protein (mg/dL) 247.02+614.42 52.81+51.42 0.320
Glucose (mg/dL) 40.36+29.26 59.48+32.28 0.111
Leukocyte (/mm3) 377.00+441.16 100.86+257.16 0.077
LAS (%) 20.27+8.26 1.95+3.37 0.000

TABLE 3. THE CORRELATION BETWEEN CSF EXAMINATIONS AND DIAGNOSIS

Diagnosis
CSF examinations Bacterial Non-bacterial Total (n) Fisher’s
meningitis meningitis exact test
(n) (n)
Color p=0.148
Turbid 4 2 6
Clear 7 19 26
Nonne test p=0.681
Positive 4 5 9
Negative 7 16 23
Pandy test p=0.645
Positive 8 12 20
Negative 3 9 12
Protein (mg/dL) p=0.703
>40 8 13 21
<40 3 8 11
Glucose (mg/dL) p=0.053
<40 7 5 12
>40 4 16 20
Leukocyte (/mm3) p=0.266
>1,000 2 1 3
<1,000 9 20 29
Bacterial culture p=0.000
Positive 11 0 11
Negative 0 21 21
LAS (%)
>12.35 11 0 11 p=0.000
<12.35 0 21 21

>15.0 9 0 9 p=0.000
<15.0 2 21 23

ing to our equation, the cut off value of LAS for the
diagnosis of bacterial meningitis was 12.35%.
Discussion
In this study, we excluded neonates and we did not
perform any culture for Mycobacterium tuberculosis,
fungi, and virus; we just categorized the diagnosis as
bacterial and non-bacterial meningitis.
Bacterial meningitis was found more frequently
in female than male and most commonly at the age of
<2 years. These findings were similar to those reported
by Klein et al.1 Vince et al17 reported that bacterial
meningitis might occur in all ages, but most frequently
in the first year of life.
We did not find bacteria in all Gram staining of
CSF. According to literatures, Gram staining may re-
veal bacteria in about 70% of bacterial meningitis, but
it may show no bacteria if the number of bacteria is
less than 104–105 organism/ml although the culture
shows bacterial growth.18
The etiologies of bacterial meningitis in our study
were Enterobacter aerogenes and Staphylococcus epidermidis

Paediatrica Indonesiana, Vol. 44 No. 3-4 • March - April 2004 • 63

 Paediatrica Indonesiana
(3 patients each), Alkaligenes faecalis (2 patients),
Citrobacter difersus, Acinetobacter, and Serratia marcesen
(1 patients each). Wenger et al 19 reported that the eti-
ologies of bacterial meningitis were Enterobacter (9%),
Serratia (9%), S. epidermidis (8%), and Acinetobacter (6%).
Michelow et al16 reported that compared to LAS,
other CSF examinations (leukocyte count, Gram
staining, and culture) and blood culture were less sig-
nificant in diagnosing bacterial meningitis. In this
study, there was no significant correlation between
the color of CSF, Nonne/Pandy test, protein, glucose,
leukocyte count and the diagnosis of bacterial men-
ingitis. It means that those examinations might not
be used as a single diagnostic tool without any addi-
tional examinations.
There were only several hundreds of leukocytes
per mm3 CSF in the early stage of bacterial meningi-
tis, even mononuclear cell may be more prominent.20
On the other hand, the CSF of non-bacterial menin-
gitis, such as viral or aseptic ones, in their early stage,
may be dominated by PMN.20-22 Polk et al23 reported
several cases of bacterial meningitis with normal leu-
kocyte count in their CSF.
Protein level in CSF may vary depending on age,
number of cell, etiology, and damaged cells. High pro-
tein level may be found in either bacterial meningitis or
other conditions.22 Thus, we can not confirm the diag-
nosis of bacterial meningitis based on protein level only.
CSF glucose level may be low in tuberculous
meningitis or if sample is examined in more than 90
minutes after LP due to the consumption of glucose
by bacteria for their metabolism. The level of glucose
in CSF is influenced by blood glucose level.24 Thus, it
is better to examine both blood and CSF glucose level,
and then determine their ratio. It was not done in our
study because of our limitation.
There was a very significant correlation between
LAS and the diagnosis of bacterial meningitis
(p<0.001). Based on these data and by regression
analysis model, we found the cutoff value of LAS was
12.35%. Garty et al reported that the cutoff value of
LAS was 15%.15
In our study, we found that with the cutoff value
of LAS at 12.35%, the sensitivity and specificity were
100% each. Meanwhile, with the cutoff value at 15%,
they were 81.8% and 100% respectively. Michelow et
al16 reported that the higher the cutoff value, the lower
the sensitivity is. They showed the sensitivity of LAS
64 • Paediatrica Indonesiana, Vol. 44 No. 3-4 • March - April 2004
was 92.5%-98.5% and its specificity was 64.3%-88.1%.
Garty et al15 reported the sensitivity and specificity of
LAS to diagnose bacterial meningitis were 88%-94%
and 100% respectively.
We concluded that there was a very significant
correlation between either LAS or CSF culture and the
diagnosis of bacterial meningitis. Therefore LAS may be
used as a fast and simple alternative diagnostic tool to
confirm the diagnosis of bacterial meningitis.
References
1. Klein JO, Feigin RD, Mc Cracken GH. Report of the
task force on diagnosis and management of meningi-
tis. Pediatrics 1986;78:959-79.
2. Snyder RD. Bacterial and spirochaetal infections of the
nervous system. In: Swaiman KF, editor. Pediatric neu-
rology: principles and practice. Toronto: Mosby Co;
1994. p. 611-43.
3. Berkley JA, Mwangi I, Ngetsa J, Mwarumba S, Lowe
BS, Marsh K, et al. Diagnosis of acute bacterial menin-
gitis in children at a district hospital in Saharan Afri-
can. Lancet 2001;357:1753-7.
4. Rutledge J, Benjamin D, Hood L, Smith A. Is the cere-
brospinal fluid lactate measurement useful in the man-
agement of children with suspected bacterial meningi-
tis? J Pediatr 1981;98:20-4.
5. Coral CJ, Pepple JM, Moxon ER, Hughes WT. C-reac-
tive protein in spinal fluid of children with meningitis.
J Pediatr 1981;99:365-9.
6. De Beer FC, Kirsten GF, Gie RP, Beyers N, Strachan
AF. Value of C-reactive protein measurement in tu-
berculous, bacterial and viral meningitis. Arch Dis
Child 1984;59:653-6.
7. Gutteberg TJ, Flaegstad T, Jorgensen T. Lactoferrin,
C-reactive protein, á-1 antitrypsin and immunoglobu-
lin GA in cerebrospinal fluid in meningitis. Acta Pediatr
Scand 1986;75:569-72.
8. Nadal D, Leppert D, Frei K, Lamche H, Fontana A.
Tumor necrosis factor-á in infectious meningitis. Arch
Dis Child 1989;64:1274-9.
9. Mustafa MM, Ramilo O, Llorens XS, Olsen KD,
Magness RR, Mc Cracken GH. Cerebrospinal fluid
prostaglandins, interleukin 1-â and tumor necrosis fac-
tor in bacterial meningitis. Clinical and laboratory cor-
relations in placebo treated and dexamethasone treated
patients. Am J Dis Child 1990;144:883-7.
 Meitha PE Togas et al: Correlation between leukocyte aggregation score of CSF and bacterial meningitis
10. Lopez-Cortez LF, Cruz-Ruiz M, Gomez-Mateos J,
Jimenez-Hernandez D, Palomino J, Jimenez E.
Measurement of levels of tumor necrosis factor-α
and interleukin 1-α in the cerebrospinal fluid of
the patients with meningitis of different etiolo-
gies: utility in the differential diagnosis. Clin Inf
Dis 1993;16:534-9.
11. Ramilo O, Mustafa MM, Porter J, Saez-Lorens X,
Mertsola J, Olsen KD, et al. Detection of interleukin
1-α but not tumor necrosis factor-α in cerebrospinal
fluid of children with aseptic meningitis. Am J Dis
Child 1990;144:349-52.
12. Dulkerian SJ, Kilpatrick L, Costarino Jr AT, Mc Cawley
L, Fein J, Corcoran J, et al. Cytokine elevations in in-
fants with bacterial and aseptic meningitis. J Pediatr
1995;126:872-6.
13. Shimoda BK, Okamura S, Omori F, Mizuno Y, Hara T,
Aoki T, et al. Granulocyte colony-stimulating factor in
cerebrospinal fluid from patients with meningitis. Blood
1991;77:2214-7.
14. Bhisitkul DM, Hogan AE, Tanz RR. The role of bac-
terial antigen detection tests in the diagnosis of bacte-
rial meningitis. Pediatr Emerg Care 1994;10:67-71.
15. Garty B, Berliner S, Liberman E, Danon YL. Cere-
brospinal fluid leucocyte aggregation in meningitis.
Pediatr Inf Dis J 1997;16:647-51.
16. Michelow IC, Nicol M, Tiemessen C, Chezzi C, Pettifor
JM. Value of cerebrospinal fluid leucocyte aggregation
Paediatrica Indonesiana, Vol. 44 No. 3-4 • March - April 2004 • 65
in distinguishing the causes of meningitis in children.
Pediatr Inf Dis J 2000;19:66-72.
17. Vince J, Biddulph J. Infections of the nervous system.
In: Jeliffe DB, editor. Child health in the tropics. A prac-
tical handbook for health personnel. 5th ed. USA: ELBS;
1985. p. 148-55.
18. Hashimoto T. Gram stain technique [citated on Au-
gust 13, 2002]. University Health System. 1995-
2000. Available from URL: http://www.meddean.
luc.edu/lumen/Deptwebs/microbio/med/gram/
tech.htm.
19. Wenger JD, Hihgtower AW, Facklam RR, Gaventa S,
Broome CV. Bacterial meningitis in the United States
1986: report of a multistate surveillance study. J Inf Dis
1990;162:1313-23.
20. Weil ML. Infections of the nervous system. In: Menkes
JH, editor. Textbook of child neurology. 4th ed. London:
Lea and Febiger; 1990. p. 327-45.
21. Hutchison JH, Cockburn F. Practical pediatric problems.
6th ed. London: Llyod-Luke; 1986. p. 526-34.
22. Haerer AF. De Jong’s neurologic examination. 5th ed.
Philadelphia: JB Lippincott Co; 1992. p. 777-88.
23. Polk DB, Steele RW. Bacterial meningitis presenting
with normal cerebrospinal fluid. Pediatr Inf Dis J
1987;6:1040-2.
24. Menkes JH. The causes of low spinal fluid sugar in
bacterial meningitis: another look. Pediatrics
1969;44:1-3.